Thunderbolt Successor to Boast of 50 Gbps Bandwidth

[btarunr]

For the greater part of the last decade, PC device connectivity was limited to the 480 Mbps bandwidth of USB 2.0. The pressing need for more bandwidth to run external hard drives and disk racks was alleviated by eSATA, but eSATA lacked the versatility of USB. After quite some delay, came the next big version of USB, the USB 3.0 SuperSpeed, with its massive 5 Gbps bandwidth, plenty for fast and capacious flash drives, and external storage devices.

There was, however, a potential bottleneck lurking with running SSD-based RAID boxes in USB 3.0, as many SATA 6 Gbps SSDs are getting close to the bandwidth limit of USB 3.0. There has also been the need for an interconnect faster than USB 3.0 for high-bandwidth applications such as lossless ultra high definition video streaming in professional environments, and hence came Thunderbolt, which is a copper-electric variant of a fiber-optic interconnect Intel had been working on, codenamed Light Peak. The successor to Thunderbolt is reportedly already under development at Intel Labs.

Thunderbolt delivers 10 Gbps of bandwidth over copper wire, but there's no guaranteeing its market longevity with the ever-increasing demand for bandwidth with applications in the future. As early as 2015, Intel will have developed a new device interconnect standard to replace Thunderbolt. The new interconnect will be able to deliver a [currently] mind-boggling bandwidth of 50 Gbps over distances as long as 100 m. The announcement came from Jeff Demain, strategy director of circuits and system research at Intel Labs, at a company event in New York.

Thunderbolt is able to make use of its 10 Gbps bandwidth to drive high-bandwidth video encoding applications in environments with external storage, as well as connect high-resolution displays over the DisplayPort protocol. The future 50 Gbps interconnect will build on Thunderbolt's applications by upscaling the bandwidth. There is, however, no definitive word on whether the future interconnect will maintain any kind of compatibility with Thunderbolt. "We see them as complementary. It's the evolution of these connectors and protocols as they move forward," Demain said.

It is likely that Intel will have developed silicon photonics to a greater degree by 2015. At least it should be able to put optical transmitter and receiver into a single chip, small enough to be fitted into smartphones and tablets.

Source: PCWorld

[thunderising]

nobody will pay for a 300$ processor to use your lousy thunderbolt..

apple mac fanbois to death

[slyfox2151]

Originally Posted by thunderising

nobody will pay for a 300$ processor to use your lousy thunderbolt..

apple mac fanbois to death

$300? thats very cheap to anyone who would need this kind of speed..... your argument is invalid, plenty of people would pay more.

[DanTheBanjoman]

Originally Posted by slyfox2151

$300? thats very cheap to anyone who would need this kind of speed..... your argument is invalid, plenty of people would pay more.

Amen. I'd gladly pay that for that bandwidth if I can transport IP over it.

[Bundy]

Originally Posted by thunderising

nobody will pay for a 300$ processor to use your lousy thunderbolt..

apple mac fanbois to death

WTF you talkin about?

On topic : sounds very appealing and I'd love to see those sort of speeds sooner.

[LAN_deRf_HA]

Originally Posted by thunderising

nobody will pay for a 300$ processor to use your lousy thunderbolt..

apple mac fanbois to death

It's like you said the most random thing you could think of. People will pay $300 for an Intel cpu because of it's performance, whether it has thunderbolt or not is irrelevant. Having it certainly doesn't hurt anything.

More on topic, does anyone know if this has the latency response to replace sata? Cause even sata 6 woefully slow. We're only a few ssd gens away from a 10x speed bump and sata specifications are updated far too infrequently.

[FordGT90Concept]

I thought Light Peak was supposed to be on the market this year. Instead, we got copper-based Thunderbolt. Now Intel is saying we have to wait until 2015 for Light Peak and there's no promises even that will be fiber. I was getting all excited over the possibility of fiber taking over as the primary system interconnect and now, we're pretty much talking about an improved CAT7a networking cable which has already been done. How disappointing.

I think I'll file Light Peak in the same pile as Larrabee. Big on talk: short on delivery.

[1c3d0g]

What I hate about these new technologies is just how LONG it all takes for these damn things to be implemented. I needed this YESTERDAY, Intel! Let's get it going already!

[inferKNOX]

Despite it's "electric shock" symbol, it doesn't look like T/Bolt (or at least the 50Gbps one) will deliver power across the cable since it's fibre. Seems like T/Bolt will be relavent to data-only links.
In that case USB & Ethernet (with PoE) will remain quite relavent because of their ability to power the device across the cable.

OT: Looks like Intel is with the illuminati, LOL!

[happita]

I think I'll keep my portable USB 3 external HDD, but thanks though Intel for trying new things and getting the industry to think about innovation!! I'll wait until Light Peak comes out, USB 3 more than suits my needs for now.

[btarunr]

Originally Posted by inferKNOX

Despite it's "electric shock" symbol, it doesn't look like T/Bolt (or at least the 50Gbps one) will deliver power across the cable since it's fibre.

Thunderbolt is an electric cable, not optic.

[v12dock]

SATA 6 Gbps SSDs are getting close to the bandwidth limit of USB 3.0

How long did USB 2.0 last after SATA passed it in speed...

[btarunr]

Originally Posted by v12dock

How long did USB 2.0 last after SATA passed it in speed...

Very, very long. Blame whoever was behind USB 3.0 for it. The fact that eSATA came to be shows that the gap between USB 2.0 and USB 3.0 was a failure, and does not in any way validate USB 3.0 having the same amount of market time.

[digibucc]

Originally Posted by v12dock

How long did USB 2.0 last after SATA passed it in speed...

yeah but usb is used almost exclusively externally, whereas sata is almost exclusively internal.
even esata isn't used for anything other than storage. they catered to different segments so
it's no wonder usb stuck around.

speed is not the most important factor believe it or not. overhead, component cost, market
ubiquity. simply being around long enough allows a technology to become (a) standard.

Originally Posted by btarunr

Very, very long. Blame whoever was behind USB 3.0 for it. The fact that eSATA came to be shows that the gap between USB 2.0 and USB 3.0 was a failure, and does not in any way validate USB 3.0 having the same amount of market time.

you're talking as though usb 2.0 is not sufficient in speed for 90% of users, which it still is.
sure faster is better, but most people don't care. the majority of consumers will look at
2.0-3.0 as a worthless upgrade, especially until devices start making use of it. then they
have to buy new external devices, to gain ANY benefit from 3.0... I know you are the
resident hardware dude... but i don't think 2.0 is so far over-extended, except in a strictly
technological sense.

[btarunr]

Originally Posted by digibucc

you're talking as though usb 2.0 is not sufficient in speed for 90% of users, which it still is.

USB 2.0 is insufficient speed for 99.99% of users who are using it for any form of data transfer. It has sufficient bandwidth for most other applications (such as driving printers/scanners/input devices).

[digibucc]

Originally Posted by btarunr

USB 2.0 is insufficient speed for 99.99% of users who are using it for any form of data transfer. It has sufficient bandwidth for most other applications (such as driving printers/scanners/input devices).

agreed there personally - not for the percentage.

if it were measured by what is possible, yes it's inordinately slow. but i highly doubt 99% of
users/consumers are using their usb 2.0 hard drives, and being put off by the speed. i'm sure
they'd be happy with faster, but insufficient to me means being to poor for use, which i don't
agree that it is for most users.

i don't think they think much about it, it just is.

[v12dock]

Originally Posted by btarunr

Very, very long. Blame whoever was behind USB 3.0 for it. The fact that eSATA came to be shows that the gap between USB 2.0 and USB 3.0 was a failure, and does not in any way validate USB 3.0 having the same amount of market time.

But how many people needed transfer over 40MB/s back in 2003 onto an external device

[btarunr]

Originally Posted by digibucc

agreed there personally - not for the percentage.

if it were measured by what is possible, yes it's inordinately slow. but i highly doubt 99% of
users/consumers are using their usb 2.0 hard drives, and being put off by the speed. i'm sure
they'd be happy with faster, but insufficient to me means being to poor for use, which i don't
agree that it is for most users. i don't think they think much about it, it just is.

99% is a safe bet figure for the percentage of external storage devices (flash drives, external hard drives) that use USB 2.0, and hence have insufficient amount of bandwidth. USB 2.0 gives you a measly 60-odd MB/s bandwidth, which is circa 1996 (when IDE with UDMA-3 gave you 66 MB/s).

It took too long before USB 3.0 came to be, and established some sort of parity with internal storage standard prvealent for the day. (USB 3.0 and SATA 6 Gbps).

Originally Posted by v12dock

But how many people needed transfer over 40MB/s back in 2003 onto an external device

I don't know, but certainly a LOT of people needed to transfer over 40 MB/s in 2005, 2006, 2007, and 2008 (when there was still no USB 3.0 in sight, and when hard drives would easily cross 100 MB/s).

[Imsochobo]

Originally Posted by btarunr

Very, very long. Blame whoever was behind USB 3.0 for it. The fact that eSATA came to be shows that the gap between USB 2.0 and USB 3.0 was a failure, and does not in any way validate USB 3.0 having the same amount of market time.

Intel was the leader of usb3 that everyone just went away from and made their own if i remember right, but intel is the one to blame, to make sure its not adopted widely for use of thunderbolt ?

who knows, intel breaks and makes

[inferKNOX]

Originally Posted by btarunr

Thunderbolt is an electric cable, not optic.

Are you talking about the current itteration or the 50Gbps one to come?
I'm referring to the latter.

[digibucc]

Originally Posted by inferKNOX

Are you talking about the current itteration or the 50Gbps one to come?
I'm referring to the latter.

well i assume the one to come will have a different name, no?

[inferKNOX]

Originally Posted by digibucc

well i assume the one to come will have a different name, no?

Well, based on FordGT90Concept's comment, I was assuming it's one and the same, with the current itteration merely being a improvisation and the Light Peak being the true Thunderbolt?

[tigger]

It seriously makes me laugh when i see 2TB usb2 drives, how long would it take to dump a TB of stuff on it, at least an hour i reckon. Usb2 is now only for peripherals and not usb hdds. Now a usb3 external hdd thats a differant matter

[Imsochobo]

Originally Posted by tigger

It seriously makes me laugh when i see 2TB usb2 drives, how long would it take to dump a TB of stuff on it, at least an hour i reckon. Usb2 is now only for peripherals and not usb hdds. Now a usb3 external hdd thats a differant matter

we were already at usb2 limits in 2001 I think, I'm pretty sure my two 45gb ibm bombtimers did those speeds.

[thunderising]

LOL I was just trolling, like

"I'm 1st to comment" and "Intel your pricing Sucks" kinda post..

but ended up with random non-sensical gibberish..

what really made my day was reading you guys reply, top class +1000